I think the change looks OK.
I agree that this case is unlikely to come up in the real world, so no
regtest seems acceptable; tag the bug with noreg-hard.
(Another option might be a test in a seldom-run Tier that @requires a
large amount of heap.)
-Brent
On 5/11/19 3:07 PM, Ivan Gerasimov wrote:
Hello!
An integer overflow during array size calculation can happen in a case
of loading extremely huge class file (which is unlikely in the real world).
It is possible to create a regression test (see the bug), though I doubt
it would carry much weight while requiring much memory.
I did check manually that the POC runs fine with the patched JDK.
Would you please help review the fix?
BUGURL: https://bugs.openjdk.java.net/browse/JDK-8223730
WEBREV: http://cr.openjdk.java.net/~igerasim/8223730/00/webrev/
